also, this is the 316$ part, not the 1000$ one, so the extra 900Mhz of the XE might help a little :D .
Printable View
also, this is the 316$ part, not the 1000$ one, so the extra 900Mhz of the XE might help a little :D .
... or take a quick look at the Orb?Quote:
Originally Posted by nCrusader
http://img456.imageshack.us/img456/9106/cputest25jd.png
sweet results but damn 21 pages of posts in 29 or so hours is crazy i cant keep up lol
Quote:
Originally Posted by nCrusader
CPU scores in this 3D tests dont mean anything. Heck I have had the Dothan at 2.6Ghz ~800FSB and the CPU score was higher than when it was at 2.8Ghz ~930 FSB....
Alex
These are very promising results considering it's all stock and on buggy hardware/software.
Thanks for the results Victor
Quote:
Originally Posted by nCrusader
thats strange I was testing my system with my Gskill 3-4-4-8 timings which suck but my redline are fried :(
I was hitting 410/762 @ 2.86 ghz with ram at 260 mhz 3-4-4-8 *11
the revised conroe #s seems inline at 398/738 but if you take 798/389 that looks to be a multi of 2.05 which is impossible isn't it???? and my 762 score took at near 34 secs to complete too.
Just wondering if having 4 threads going makes that CPU jump in eff???
The Conroe seems to have about the same 1.85 effective increase as an AMD does with two cores vs one core in cinebench.
These results are going to kill AMD sales. In fact, its going to hurt Intel sales for the next few months, as everyone is going to want to wait for Conroe! Though I have to say I'm very happy with my Pentium D 920 (Presler). And I've seen some very fast CedarMills too.
I'm going buy some Intel stock as its at a 3 year low. By Q4 the results should be impressive.
You are talking about cinebench 2003, not 9.5 (2003 is slower than 9.5) and VictorWang has tested 9.5 version.Quote:
Originally Posted by iboomalot
738 with Conroe @2.4 is a bit faster than a X2 @2.6 in Cinebench 9.5:
http://www.3dfluff.com/mash/cb95/top.php
Look here an oced X2 3800+ @2.6: 709
as an AMD user and I have been disappointed at the P4 screwup by Intel I will have to say if the result so far from a 2.4 ghz Conroe is very impressive.
My 4400+ is basically a FX-62 when running at 2.8 ghz and Conroe is equal to that for a price of $360.00ish and is running at stock speed when my 4400+ has used nearly 50% of its overclocking headroom to catch up to the conroe.
This will be alot of pressure on AMD to ramp up some kind of equalizing CPU and if Conroe doesn't have a cold bug and can OC like the yonah cores then this should be interesting to say the least.
I haven't used an Intel chip in 6 yrs but I will be looking at a Conroe and Phase change if no cold bug appears and Intel doesn't do some kind of OCing restrictions on the chips.
Its about damn time Intel finally stepped up to the bat and maybe has hit a home run finally.
Quote:
Originally Posted by PetNorth
Oh my bad I will have to retest using the new bench
either way still impressive.
Hmm is it me or are there 2 diferent 3DMark05 CPU scores there? One with 6k and one with 8k?
Quote:
Originally Posted by Dragex
Hmmm.......I think "it's you"........:D
One is PCMark05 and the other one is 3DMark05
:stick: Hahaha, ok, thanks.Quote:
Originally Posted by hipro5
VictorWang, can you run WinRAR 3.6 beta 1 built-in benchmark? (it's multithreaded)
http://www.rarlabs.com/download.htm
thanks!
is 1131 a good speed on that winRAR bench??? system at 2.7ghz
and got 1200 flat @ 2.85 ghz
I redid my setup cinebench to match the 2.4 conroe was 2.705 ghz with ram at a slow 245 mhz 3-4-4-8
still impressive compared to older intel CPUs
a bit low I'd say, my X2 @2.2 get about 1025.
This is so fricken nuts...I can't wait to see more results once software/hardware has had a chance to catch up and these chips are really able to stretch their legs. :cool:
why would an intel processor have a cold bug? its the IMC on the AMD's that make them do that...
conroe is extremely impressive. and you're using a 6600le, like me! haha i hate this thing :(
There must be some bug or (again?) im reading something wrong, look what I get with my CPU on 3DMark05... odd.
http://img218.imageshack.us/img218/1...core4cz.th.jpg
Compared to his:
http://vic.expreview.com/attachment/...05+cputest.jpg
Dude, he fixed the test, his score is 8,320
http://www.xtremesystems.org/forums/...&postcount=505
i think i just creamed myself.
what nice pi results. :D
yonah/merom/conroe share the exact same L2's (design wise and latency wise) according to anand's coverage of IDF, so it should be 14ns for the full 4MBQuote:
Originally Posted by Willis
god i cant wait for some real results from a real mobo with OC options, better RAM, and extreme cooling (Fuggers cascade should provide us with that).
Also, I,d like to see not much MORE than real results. After following this thread for a few days, ive seen maybe 20 relevant posts. The rest are like, "OMG!!1!! thats amazing can you run ____(some damn benchie)?"even after Victor has said hes having stability problems.
"wow that looks good, but am2 will pwn intel all over again"
"can you OC?" agian, its been said... NO!
"FCG and Fugger promised results by whenever and they didnt deliver wah wah wah! Now i have to wait a few more days for results that would otherwise taken months to reach us."
Aside from my lil rant, thanks to Victor, FCG and Fugger for providing such exciting news so early :toast:
how hot is this proc??
Victor, can you run Super Pi 2m?
Regards..
yeah! what are temps like?Quote:
Originally Posted by StyM
I gotta ditch this Gskill ram 200*11 at 3-4-4-8 921 on winrar bench OUCH !!!!
oooooohhhhyeah~~~~~~~!!!
78gtx 512m no running with my lovely E6600.
3Dmark05 benching ...
exciting !!!
Good to hear :) Keep us informed. :toast:Quote:
Originally Posted by VictorWang
Just curious what did you do to get the 7800GTX to work if you don't mind me asking?
Oh my god i cant wait :DQuote:
Originally Posted by VictorWang
Keep up the awesome work Victor!
well, whats the deal?Quote:
Originally Posted by VictorWang
Very exciting news indeed. Waiting impatiently for the 7800 512mb results.
and put some 3dmark06 on in...stat!
vic, you get the 2x1g stix to work yet?
On Victor's page, did you guys notice how dual SPI only does a run at 25seconds? That's 4 seconds lower than in single.
Can anyone on a DC AMD do a single and then a dual SPI run and see what the difference in times is? (make sure you seperate both instances to run on one core only by setting affinity)
naturally, it will be this way because of the shared cache. oh well.. 4mb is still a ton and 25s at stock is farking crazy nonetheless :)Quote:
Originally Posted by The Coolest
Quote:
Originally Posted by VulgarHandle
1G*2 still cannot run, only 256m*2 running.
xfx 78gtx 512m on stock.
580_1730, LOD=0
3Dmark05=10517
CPUmark=8621
put some 3dmark06 on it!Quote:
Originally Posted by VictorWang
Is 10,000 3Dmark05 really very good?? :confused:
I get over 12,000 3dmark05 with the rig in my sig no problem
3dmark05 is a graphics card benchmark =[ You have a x1900xt
X1900 Runs 3DMark05 better than 7800GTX, your rig is at 2.9GHz, this is at 2.4, his ram is only at DDR533 and 4-4-4 timings, yours is higher and tighter than that, plus you have 1GB he is working with 512.Quote:
Originally Posted by ceevee
All in all, very good initial results, he can't tweak anything at the moment, except the video card maybe.
Damn Conroe is killing it @ the momenet getting 10K plus with stock parts and with a 2* 256 ram.with buggy bios.Now that's somethin good job Victor:clap:
Agreed. Its too hard to read much of anything into the '05 score because of some problems that still have to be sorted through. Very unlikley the CPU is getting in the way of the rest of the system, no doubt its the other way round.Quote:
Originally Posted by RyderOCZ
3D benches that allow the members to project OC'ed results, will make for a dramatic finish to this initial testing thread. :D
Let's see some real benchies, like 3dmark01 :D
and an ATI cardQuote:
Originally Posted by wwwww
05 CPUmark=8621 is almost good enough for 1st page of orb (edit:@ stock)
The X2@3001 MHz/1gb/x1800 on the first page hits CPU Score: 8662
LOL :drool
If this is restricted by driver/bios issues... wow.
The geeks of earth need some sleep...we need some sub zero temp runs on a good board... on second thought that might make things worse.
the 3dmark05 score is phenomenal. I don't know much about 3dmark tweaking... but my 7900GT clocked higher than his 7800GTX (mine is clocked at 600/1860) with my 170 @ 3.15 Ghz benched exactly 10,500 on 3dmark05. Victor actually beats my system here even though my GPU is clocked higher and my 170 is 58% overclocked.
So from what i've seen 2.4 Ghz Conroe is faster than my 170 @ 3.15 Ghz
WOWZA
He has the 512mb version
Again, it is a gpu test. Cpu would be 2001SE
To compare (kinda) an AMD 4000+ (2.4GHz 1MB L2 cache) with 7800 GTX gets 3Dmark05 video score of 9395 (corrected) http://www.xbitlabs.com/articles/vid...900xtx_37.html (:eek: )...Quote:
Originally Posted by VictorWang
Sure, 4000+ is a single core, but from what I've seen 3dmark05 doesn't gain much from dual cores.
Not too bad if you ask me. :toast:
Mike
PS
Current AMD 4000+ price = ~$330 >>> same price Conroe 2.4 is to be released at.... (AMD-roids, keep your panties on, this is just a simple comparasent of prices that will most likely change, on two different classes or CPUs, but then again, single core, double core, it's all the same to 3dmark05)
Yeah 3Dmark05 must be much more GPU dependent. That is the only thing that can explain why I can beat Absolute's scores above so handily when we have the same CPU and his is clocked 115Mhz higher...
Actually my highest 3dmark05 score was 12,600 :D
guess the X1900XTX pwns... :slapass:
I hope its the fact that your running only 512. Or hopfully that its buggy as heck. I scored a 11170 in 3dmark05, but with 2 gb of memory. See my Signature.Quote:
Originally Posted by VictorWang
http://service.futuremark.com/compare?3dm05=1884411
Here is mine to compare with, stock 7900GTX(which shouldnt really very from the 7800GTX 512. (650mhz core, 1600mhz Memory) I think this thing kicked ass in the SuperPI test, but isnt doing so well in 3Dmark.
~Mike
its def gpu dependant
01 is where its at..
looks like 9500 - 9900 is the score for a X2 4800+ with a GTX 512 and 2gb of ram... 800 more points is pretty damn good for a gpu benchmark.
http://service.futuremark.com/compare?3dm05=1761119
Guys, guys, guys... the GPU dependency means that any increase in score due to the CPU represents a very large CPU power!
About the 512mb thing, 512mb versus 256mb makes 0 difference in 3dmark
And the score being higher because it's buggy? Buggy can only make scores lower, not higher.
Quote:
Originally Posted by arisythila
622_1880 3Dmark05= 11225 :p: only 256m*2 4-4-4-12
if 79gtx @ 650_1800 with 1G*2 ram , at least 12000 :D
Wow that is a good CPU score. I just ran 3dmark05 again right now on my machine (no tweaking just spur of the moment) and got:Quote:
Originally Posted by VictorWang
http://img82.imageshack.us/img82/4008/ma059hq.jpg
So Conroe is defintiely pwning AMD dual core current gen!
i'm confused i asked one of my contacts at a mobo company and he told me that the only mobos that will support conroe officially i guess are:
965 and 946 series chipset
any one can explain this to me?? also he told me that by august they will have the mobos ready
Nice Victor, keep it coming!
well, hes using 965 for these tests, and 975 is supposed to work from everything i've heard, but so far no current boards are working with Conroe that we've seen.
:eek: :eek: :eek:Quote:
Originally Posted by VictorWang
Getting better and better... Again, let me emphasize, this is the kinda performance you'll be getting from a ~$300 CPU at STOCK speeds...:woot:
Loving it.
Mike
PS
Great job Victor :toast:
Keep up the good work :DQuote:
Originally Posted by VictorWang
any info on the cpu temp??? idle ??? full load ???
Some very nice scores VW, thank you for sharing it with us. At stock this CPU is really killing almost everything. Hope to see a 975X board + this chip go for Rampage.
Looking here http://www.xbitlabs.com/articles/vid...900xtx_37.html, I figured the performance ratio (for 3Dmark05 only!!!) of ATI 1900XTX to nVidia 7800GTX 512Mb cards is about 1.11, sooooooo if you had a 1900XTX you'd be getting ~12460 3Dmark05 score... :eek: :banana: :eek:Quote:
Originally Posted by VictorWang
I forsee AMD hardware getting very cheap in the coming months...;)
Mike
cant wait to see this thing overclocking, a core based on low power principles when given lots of high power must be crazy.
and no cold bug either
Conroe will probably be the chip that I switch to phase change for...
Thanks for all the great results Victor, any info on temps??
Quote:
Originally Posted by ceevee
just a heatsink with 2 heatpipes inside.
no fans at all :p: :p:
the HS just a little bit warm, maybe 25c (ambient=22c)
some results updated, welcome to my blog:
http://vic.expreview.com/index.php
So overall it's probably pretty cool running? Wonder if you could get any temp probes in there, no temp support from the motherboard?
ya, cool like yonah (maybe a little bit cooler......since lower vcore)Quote:
Originally Posted by Absolute_0
no temp monitor function in the mobo :(
Are you running passive cooling on that CPU? That's insane!
damn it VW i want to see this thing running past 3GHz.....common dude show us the goods :)
yes..... only oced@3 GHz :-)
cause yonah just hit the 17sec 1m
and an impressive 16 min 32 m
:clap: :clap: :clap: :clap: :clap: :clap:
Geez, what are we gonna do w/all those 120.3 rads! :eek:Quote:
Originally Posted by VictorWang
Strat
all [Cinbench&PCMark&other] tests???Quote:
Originally Posted by VictorWang
I got 11200 with 3Ghz dothan and gtx @ 650/770 (modded bios, no throttle etc)
If I remember right 10Mhz on core gave like ~90 3dmarks
I got 11200 with a 2.82GHz Toledo, GTX256MB @ 600+50/1670, modded BIOS (*but*, with GTX-512 timings it would need 600+50/1785 to get the same score, for some perspective [remember it is 256MB]).
Considering how high of a CPU score that is, I feel this board is seriously hampering 3D performance, or something is (possibly the RAM, IDK). So, in case anyone is wondering, I see these scores as around 5-10% lower than they will be....
why is there no fx-60 on the benchmarks???
Good question ;)
Ah...God almighty...can you guy's smell that??? Yes it's the return of Intel. Like I said in another thread ; I have much love and respect for AMD. It's all I used for the past three-four years after both Northwood lines. But damn, Intel sure can create a damn good cpu when they put some effort in. :)
Welcome back Intel. :toast:
Is it possible to do the pin mod on Conroe to achieve higher fsb and higher clock? May already been disscused.
No better words were spoken in this thread apart from Victor, so very true. :fact:Quote:
Originally Posted by LOE
I dont understand people who are anti-Intel, anti-Conroe or whatever for the sake of it.
If Conroe is the fastest thing around when its released, which it certainly looks like it will be, then I don't care whether its Intel, AMD or Von Dutch manufacturing it.
FWIW I have an Intel P4 at the moment, was considering an AMD X2/FX, and now probably waiting for Conroe. I have no allegiances either way, allegiances to chip manufacturers is just cutting your own nose off to spite your face.
3dmark 2001 please ? :)
Regards
Andy
Hey, forums are back ;)
Damn people can complain ! Oh well..Intel guyz will be sitting with a faster cheaper system in the end of the day so whine all you like.
Im glad to see the Forums are back online!
Victor ppl wont believe ur running your conroe without a fan on that heatsink. Heck what more do they want ? Do they want to stand next to you to see the results :stick: lol Dont worrie I believe ya, you have no reason to lie.
If it wasnt for some of your hardware being under NDA we would have seen pictures of his rig. But people just dont understand it :(
ORCBEAST
How is it going with the bios issues? Vic's bios is bugging and FCG and FUGGER doesn't have a bios which supports Conroe? (yet) I hope they get everything up 'n running!
Apparently the weakness of the Intel platform is it's FSB (or whatever the Hypertransport counterpart is), and it'll constrain the performance of the otherwise impressive Conroe chip.
Because of this thread, i'm already setting aside my $500-$600 for that mid-year upgrade.. previously, it was supposed to be a s939 one (prices would drop when AM2 arrives)..
But after reading this thread since it started.. an E6600 with some good generic DDR2 + a good OCing mobo would be the best "play/hobby" rig that money could buy come Q3 2K6..
;)
Hope this is the right spot for this info. It seemed relevant to the current discussion as it involves info regarding mobo support for Conroe. I have a D975XBX rev 302 mobo, and the following is a condensed version of an e-mail exchange I had with Intel support.
My support inquiry:
Product: Intel(R) Desktop Board D975XBX
Product_id: 2205
Bios_version: BX97510J.86A.0807.2006.0314.1158
AA_Number: AAD27094-302
Processor_manufacturer: Intel(R)
Issue: I just purchased this top of the line board (my first Intel manufactured board), and now find that it will not support Conroe. This board was obsolete before I purchased it! This is completely unacceptable. Can I exchange this board for a rev AAD27094-304 that will support Conroe?
Intel's response:
Thank you for contacting Intel® Technical Support.
Unfortunately, Intel does not exchange for different product. A return or exchange like this could only be done through your point of purchase, depending on their own policies and warranty procedures. As the customer of your point of purchase, they owe you return or exchange based on your purchase agreement/return policy; however, this is not something Intel will be able to take care of for you or assist in mediating with. Our warranty is an exchange of a damaged part for a good part of the same type; it would not provide any upgrade or difference in features on the part.
I draw three conclusions from this exchange:
1. The 304 rev appears to be the 'official' mobo to support Conroe per this reference to 'difference in features on the part'.
2. This is a sideways admission that the 302 rev boards will not support Conroe.
3. Owners of the 302 rev have no recourse but to buy a new mobo.
Sorry to be the bearer of bad news. This may not be a difinitive take on the issue, but I would not take it as a positive for current D975XBX owners either.
Quote:
Originally Posted by jermaink
Yes you have to remember that P4 is a 200FSB QUAD Pumped = 800FSB
Conroe runs 266 Quad pumped = 1066FSB
Now taking that into account. The P4 were never far off AMD if we are talking about SPEED even with there techonology and with there onboard mem controller ;)
Intel will soon go that way and then u have to tell yourself how bloody fast that will be.
ORCBEAST
@Slurr: Imho, they shouldn't be held liable..
Because in the first place. Conroe will only be available by July at least.
And you're buying the mobo now.
What's top-of-the-line now will certainly not be so come July.
I purchased (past tense) this mobo in march following the IDF benchmarks that were completed on the D975XBX. My point is that the lack of definitive information regarding mobos that support Conroe is not good, and it appears that current owners of this board are out of luck.
I see.. well, that's really what happens for prototype stuff..
Heck, even VW, FCG or Fugger have trouble making the Conroe run on any current motherboard.
Not a weakness on the desktop market. Conroe's two cores communicates over the shared L2 cache and direct L1->L1 data cache transfers, which is probably a magnitude faster than AMD's implementation for dual cores.Quote:
Originally Posted by jermaink
The fsb is the weak link in Conroe as it is the same in Netburst architecture, which cannot handle the full bandwidth of dual-channel DDR2 SDRAM.Quote:
Originally Posted by accord99
Conroe uses Memory Disambiguation (part of the Smart Memory Access) and a better prefetch, branch prediction, L1/L2 Cache management and deeper buffers.
The FSB1066 is not the bottleneck.
the fsb will not be an issue for conroe in terms of bandwidth cause simply conroe is more of an amd like arc, so it does not need a whole lot of bandwidth at all. merom, on the 666 bus will most likely be just as fast.